Facial Action Point Detection Using Convolutional Neural Network
Yaya Wihardi, Rahmandani, Erlangga, Dimas Saptahadi
Universitas Pendidikan Indonesia
University of Alabama at Birmingham
Abstract
Facial action point detection is a very important task in computer vision due to various application such as in facial expression recognition, facial motion capture, and another human face analysis. This research aimed to propose a deep learning approach to identify and localize facial action points in sequences of images by using convolutional neural network. The result provide an evidence that the model can detect the points with 87% of accuracy.
Keywords: Facial Action Point, CNN, Deep Learning
Topic: Computer Science